Prediction of Press-Fit Quality via Data Mining Techniques and Artificial Intelligence

• Main Authors: Rene Cruz Guerrero, Maria De Los Angeles Alonso Lavernia, Isaias Simon Marmolejo
• Format: Article
• Language: English
• Published: IEEE 2019-01-01
• Series: IEEE Access
• Subjects: Press fit; predictive model; classification; association rules; ontology schema
• Online Access: https://ieeexplore.ieee.org/document/8888164/

Determining the press-fit quality of pieces in advance is of the utmost importance because it enables the reduction of the time that is invested in the process and the prevention of material losses. High predictive accuracy is essential in a classification model; however, several studies have shown that the class category of a new instance may be insufficient information for decision making. To provide additional information to the user, this study presents a novel system that is based on a hybrid model, which, in addition, to using a classifier, extracts a set of class association rules that enable the determination of which patterns influence the new instance to belong to a class category. To select the classifier, the accuracy, recall and F-measure metrics were utilized. The rules were obtained with the Apriori algorithm to show that this knowledge is automatically represented in an ontological scheme with the objective of applying the Pellet reasoner.